The Fourth Monocular Depth Estimation Challenge
By: Anton Obukhov , Matteo Poggi , Fabio Tosi and more
Potential Business Impact:
Helps computers guess how far away things are.
This paper presents the results of the fourth edition of the Monocular Depth Estimation Challenge (MDEC), which focuses on zero-shot generalization to the SYNS-Patches benchmark, a dataset featuring challenging environments in both natural and indoor settings. In this edition, we revised the evaluation protocol to use least-squares alignment with two degrees of freedom to support disparity and affine-invariant predictions. We also revised the baselines and included popular off-the-shelf methods: Depth Anything v2 and Marigold. The challenge received a total of 24 submissions that outperformed the baselines on the test set; 10 of these included a report describing their approach, with most leading methods relying on affine-invariant predictions. The challenge winners improved the 3D F-Score over the previous edition's best result, raising it from 22.58% to 23.05%.
Similar Papers
Benchmark on Monocular Metric Depth Estimation in Wildlife Setting
CV and Pattern Recognition
Helps cameras guess animal distances in photos.
Survey on Monocular Metric Depth Estimation
CV and Pattern Recognition
Lets cameras measure real distances without special tools.
StarryGazer: Leveraging Monocular Depth Estimation Models for Domain-Agnostic Single Depth Image Completion
CV and Pattern Recognition
Makes 3D pictures from one photo.